7.23.7 [long-term] is released!

OOT don't put rcd or rcbo in series. If something happen all trip and you will face blackuot. Just saying.

I live in the countryside, and surge arresters have often saved my household equipment.
Nothing can withstand a direct lightning strike, but sacrificial surge arresters have frequently saved everything.
When I’m at home, however, I unplug everything until the storm passes.
The schema is electricy meter, fuses, discharger, my power meter, RCBO (RCD+MCB)[1] , another RCBO, etc.
The two are positioned one behind the other in the same frame,
to ensure that at least one of them, from two different brands, works.
Even though they are checked regularly, just like fire extinguishers...
Well, with life, you never know...


  1. Then it depends on the translation. In Italian is "magnetothermal and differential circuit breaker / lifesaver [salvavita]" ↩︎

There are two ways of installing RCDs:

  1. in parallel where each RCD "covers" certain circuits. If one trips, circuits behind other RCDs keep being alive. Example would be using RCD per flat (in block of flats) or per floor (in a multi-storey house)
  2. in series where all RCDs "cover" same set of circuits. Intent is to make sure that at least one of RCDs trips when need arises. This covers against failure of one RCD. And this is the way @rextended mentions of using them.

And then one can use them in combined setup: one RCD (with a bit higher residual current, e.g. 50mA) covering whole house and several RCDs (with lower currents, e.g. 10mA) to cover more "sensible" circuits (e.g. bathroom). If there's "an event" on one of "sensible" circuits, then it might be that only the "sub RCD" trips, saves person there, while general RCD doesn't trip and the rest of circuits keep being powered. Again use case by @rextended

Well ... that can be annoying. Imagine a power outage in the middle of the night ... my significant half would kill me if emergency light would come on in sleeping room. :grimacing:
In my home, I have one emergency light per floor ... which is enough to get out of house (or fetch a torch light).

damn that's a stupid typo, of course I meant I'm not installing the incoming 7.23.8, lol

Just f... great i intstalled this damm update on one of mine RB5009 and now in SSH and WinBox i got information ERROR: wrong username or password. I used those credentials few minutes earlier to start update.

Great!!!!!!! Now i need to go to another part of city and restore everything from backup via console. Great F!@#$%^& update! Congratulations!!!!!!

I tried telnet, web, winbox (old and new), ssh and nothing. I used my credentials and my collogue used his nothing. Problem is with this update and RB5009 all other routers which i has (RB760, RB4011, RB3011, E50UG, E60, L009) works fine and don't show any problems with credentials (beside that The Dude can't login to them but i can use SNMP). I need ASAP help from Mikrotik to solve this issue, because if this is urgent updated I have two choices. Updated and hope that on another RB5009 i will not have problem with credentials or keep RB5009 on 7.23.5 which works fine known that it has bug.

I'm mad because update which apparently should fix something, broke something else. I could work only with SSH but i lost completely control over router. Thanks god that this is not the most important router in organization.

Does Mikrotik check before publish this update for such advance option like login possibility after update?

I'm sorry, but that is precisely the sign... that it had already been compromised...

For starters, it’s not a 100W floodlight... just a small courtesy light...
so, if the light is on hallway, if the door is closed, you can’t see anything...
(and the mobile phones is a separate matter)
Say you’re taking a shower in the evening and the power goes out...
If the bathroom door is shut, the light from the hallway doesn't get in,
and you risk hurting yourself in the dark...
Do you see why I put one in every room?
Bright in the living room, kitchen, storage room, office, bathrooms,
soft in the bedrooms.

You base this statement on what? I have multiple routers on same public IP range and only RB5009 acting like that. Even now i took from box brand new RB5009 and put backup of one affected routers (done on 7.23.5). It was working. Update to 7.23.7 and after first successful login i can't login to it any more. Router is disconnected from WAN.

Two options.

  1. Routers are affected via dark side of force wirelessly
  2. RB5009 is affected by this update.

RCD's do not discriminate in the way fuses or MCB's do. If you have a 40A fault on a 5A circuit fed from a 40A circuit, you can generally rely on the 5A fuse or MCB clearing the fault without affecting the 40A fuse or MCB. There are whole books of fuse curves and rules for ensuring this happens.

But for RCD's it is different. If your 5A fuse circuit has the 10mA RCD and the 40A fuse circuit has a 50mA RCD and you get a 1A earth leakage fault on the 5A circuit, either of the RCD's or both may trip. To enure discrimination, it is usual for the upstream RCD to have a time delay.

Off-topic, I know, but I think it illustrates that excessive protection does not always yield an optimum result.

The way @rextended is using them (in series) I don't think he actually cares about which RCD trips ... or if both. They are mounted in the same switch box and if one (or both) trip, he can go and reset whichever tripped.

As to 50mA vs. 10mA RCD: if there's a potential leak (e.g. phase present on unearthed device surface) and some high-resistance path becomes available (e.g. human body touches the device surface), then hopefully the 10mA RCD will trip before the contact becomes conductive enough to pass 50mA (or, God forbid, multiple A). BTW, this is the reason for regulations to require low-current RCDs to be used in bathrooms ... because bathrooms offer higher risk for people (think hair dryer falling into bath tub while somebody occupies the very same bathtub). Similarly for the kitchen (e.g. electric cooktop).

And yes, when human body becomes a part of current path, 10mA v.s. 50mA makes a very big difference.

Exactly, I HOPE at least one trips...

I’d rather prefer take a walk to the electrical panel
than a walk behind, or not walk inside, the hearse...

Is this an electricians knowledge exchange or something?

Well...

if you have SMB disabled or auto/disabled and don't use containers, you should survive with 7.23.5 till next update. Or if you were fine with 7.23.6 and don't have LTE modems, go with it.

And of course no ssh, webfig, winbox etc open to unsecure networks, but I assume it is automatic for us here, whoever fails to comply with these basic rules deserves the attacks

LTS is now a joke. Great job guys. :roll_eyes:

It always was, that is why Mikrotik is calling it Long-Term without Support...

I’ve updated dozens of them from version 7.16.2->directly to->7.23.7, and not a single problem...
And that’s not all: over the years, only one version of RouterOS had a problem with passwords,
specifically when they switched from storing them in plaintext to using salted hashes... (6.29? not remember)
And the problems you're describing are exactly what happens with the honeypots I set up.
I can keep getting into them, but if I reboot them... fat chance of getting back in...

I agree this is exactly how compromised devices act, but there is also a pretty fresh CVE (well, it is a month old, but published on Sep 16) about mikrotik devices storing plaintext (i.e. readable) passwords to SPI flash. And keeping in mind how Mikrotik informs us about their fixes, we can't be sure they tried to fix it and it ended up like always (:

Cleanup done.